ix-supervised

name	ix-supervised
description	Supervised learning — regression, classification, evaluation metrics, cross-validation, confusion matrix, ROC/AUC, SMOTE resampling
disable-model-invocation	true

Supervised Learning

Train and evaluate classification and regression models with full evaluation toolkit.

When to Use

When the user asks to classify data, predict values, train a model, evaluate accuracy/precision/recall, compute confusion matrices, ROC/AUC curves, cross-validate, handle imbalanced data, or compare ML algorithms.

Algorithm Selection

• Linear Regression — Continuous target, linear relationship, OLS normal equation
• Logistic Regression — Binary/multiclass classification, gradient descent
• SVM — Binary classification with margin maximization, hinge loss
• KNN — Instance-based, no training phase, works with any distribution
• Naive Bayes — Fast classification assuming feature independence, good for text
• Decision Tree — Non-linear boundaries, interpretable rules, Gini impurity splits

Evaluation Toolkit

Metrics

• Regression: MSE, RMSE, MAE, R²
• Classification: Accuracy, precision, recall, F1 (per-class, macro, weighted)
• Probabilistic: Log loss (binary cross-entropy)

Confusion Matrix

Full confusion matrix with TP/FP/FN/TN per class, classification report (precision/recall/F1/support), and display formatting.

ROC / AUC

ROC curve (FPR vs TPR at varying thresholds) and Area Under Curve for binary classification. Use auc_score() for one-liner evaluation.

Cross-Validation

• KFold — Standard k-fold splits with optional shuffle
• StratifiedKFold — Preserves class distribution (essential for imbalanced data)
• cross_val_score() — One-liner: returns per-fold accuracy for any Classifier

Programmatic Usage

use ix_supervised::linear_regression::LinearRegression;
use ix_supervised::logistic_regression::LogisticRegression;
use ix_supervised::svm::LinearSVM;
use ix_supervised::knn::KNN;
use ix_supervised::naive_bayes::GaussianNaiveBayes;
use ix_supervised::decision_tree::DecisionTree;
use ix_supervised::traits::{Regressor, Classifier};
use ix_supervised::metrics;
use ix_supervised::metrics::{ConfusionMatrix, roc_curve, roc_auc, auc_score, Average};
use ix_supervised::validation::{KFold, StratifiedKFold, cross_val_score};
use ix_supervised::resampling::{Smote, random_undersample, class_distribution};

Examples

Confusion Matrix

use ndarray::array;
use ix_supervised::metrics::ConfusionMatrix;

let y_true = array![0, 0, 1, 1, 2, 2];
let y_pred = array![0, 1, 1, 1, 2, 0];
let cm = ConfusionMatrix::from_labels(&y_true, &y_pred, 3);

println!("{}", cm.display());
let (prec, rec, f1, support) = cm.classification_report();

ROC/AUC

use ndarray::array;
use ix_supervised::metrics::auc_score;

let y_true = array![0, 0, 1, 1];
let y_scores = array![0.1, 0.2, 0.8, 0.9];
let auc = auc_score(&y_true, &y_scores);
// auc = 1.0 (perfect separation)

Cross-Validation

use ndarray::{array, Array2};
use ix_supervised::validation::cross_val_score;
use ix_supervised::decision_tree::DecisionTree;

let x = Array2::from_shape_vec((8, 2), vec![
    0.0, 0.0, 0.5, 0.5, 1.0, 0.0, 0.3, 0.2,
    5.0, 5.0, 5.5, 5.5, 6.0, 5.0, 5.3, 5.2,
]).unwrap();
let y = array![0, 0, 0, 0, 1, 1, 1, 1];

let scores = cross_val_score(&x, &y, || DecisionTree::new(5), 4, 42);
let mean = scores.iter().sum::<f64>() / scores.len() as f64;

Resampling (SMOTE)

• Smote — Synthetic Minority Over-sampling Technique (interpolates between minority neighbors)
• random_undersample() — Reduce majority class to match minority
• class_distribution() — Diagnose class imbalance

SMOTE Example

use ndarray::{array, Array2};
use ix_supervised::resampling::{Smote, class_distribution};

// 8 legitimate (class 0), 2 fraud (class 1) — severe imbalance
let x = Array2::from_shape_vec((10, 2), vec![
    0.0, 0.0, 0.5, 0.5, 1.0, 0.0, 0.3, 0.2,
    0.8, 0.1, 0.2, 0.7, 0.6, 0.3, 0.4, 0.8,
    5.0, 5.0, 5.5, 5.5,
]).unwrap();
let y = array![0, 0, 0, 0, 0, 0, 0, 0, 1, 1];

// Before: class 0 = 80%, class 1 = 20%
let dist = class_distribution(&y);

let smote = Smote::new(5, 42);
let (x_balanced, y_balanced) = smote.fit_resample(&x, &y);
// After: class 0 = 50%, class 1 = 50%

MCP Tool Reference

Tool: ix_supervised Operations: linear_regression, logistic_regression, svm, knn, naive_bayes, decision_tree, metrics, cross_validate, confusion_matrix, roc_auc